Hazzard

My issue is more with the math of it. Since it requires holding your frames until you've got one in reserve (can't generate an in-between until you know what's next), it fundamentally makes the game less responsive.

That said, if you understand that, and like the visual smoothness of motion with more frames, then it's super cool tech. Not every game has to be treated like it's competitive Counter Strike, and I think it's really cool if you like it, but it frustrates me how poorly marketed and understood the actual technology and its compromises are.

Eh, FSR3 upscaling and FSR3 frame generation are different things. I'm personally a fan of upscaling, it's great for a sharper picture on my large 4k TV without spending a fortune on a massive GPU (I use a living room gaming PC), but not at all a fan of frame generation, as it introduces more input lag for the illusion of more frames. Not a tradeoff I'm ever willing to make, especially when VRR already does an incredible job of creating the illusion (and a degree of reality) of good performance when my framerate drops.

Sounds like a CEO who doesn't have a damn clue how code works. His description sounds like he thinks every line of code takes the same amount of time to execute, as if x = 1; takes as long as calling an encryption/decryption function.

"Adding" code to bypass your encryption is obviously going to make things run way faster.

Very much still Max's to lose. Like... 57 points? If he DNF'd twice, absolutely, Lando could take it, but even if Lando and Oscar consistently 1-2'd, with fastest lap, the most McLaren could do, if Max places 3rd every weekend it's unwinnable without sprints (11pts a weekend, x5 weekends, 55 of 57 needed points advantage).

In reality, even a decent Max Verstappen can lock out this championship with ease. So long as Red Bull don't catastrophically tank his car for a few of these weekends.

Exactly, play by the original rules, and play aggressive as all hell. You don't need almost any property, it's just fine to mortgage everything but your main set, the goal is to get one very developed set ASAP.

Not only is this a pretty effective way to win (a conservative player who lands once on a very developed property is basically out of the game), it also makes the game progress much faster, especially if other players are willing to concede before the bitter end. 2 or 3 players like this, and you've actually got a recipe for a decent time.

Eh, not much nefarious you can do by pushing data around. Taking a lot of CPU/GPU usage? Certainly, you can do a lot of evil with distributed computing. But bandwidth?

Costs a lot to host all that data to push to people, and to handle streaming it to so many as well, all for them to just... throw it out? Users certainly don't keep enough storage to even store a constant 100Mb/s of sneaky evil data, let alone do any compute with it, because the game's CPU/GPU usage isn't particularly out of the ordinary.

So not much you could do here. Ockham's razor here just says... planes are fast, MSFS is a high fidelity game, they've gotta load a lot of high accuracy data very quickly and probably can't spare the CPU for terribly complicated decompression.

Agreed, the way they can preserve the position of any object, anywhere, with thousands of objects and an obscenely large world, is exceedingly impressive.

What I don't get is why the hell any of that is a priority. It's a neat party trick, but surely 99.9% of the gameplay value of arranging items for fun could be achieved on the player ship alone.

Like... it's neat that I can pick up, interact with, and sell every single pen and fork on every table. But is it useful, with a carry weight system deincentivizing that? Fussing with my inventory to find what random crap I accidentally picked up that's taking up my weight? Is that remarkably better than having a few key obvious and useful pickups? Is it worth giving up 60FPS on console, and having dedicated loading screens for nearly every door and ladder around?

Again, it's cool that they have this massive procedurally generated world, that a player could spend thousands of hours in. But when that area is boring, does it really beat a handcrafted interesting world and narrative? What good is thousands of hours of content when players are bored and gone before 10 hours?

So like... from a tech perspective, I respect what Starfield is, and it's very impressive, but as a game it feels like a waste of a lot of very talented work, suffering from a lack of good direction at the top.

I think it is a problem. Maybe not for people like us, that understand the concept and its limitations, but "formal reasoning" is exactly how this technology is being pitched to the masses. "Take a picture of your homework and OpenAI will solve it", "have it reply to your emails", "have it write code for you". All reasoning-heavy tasks.

On top of that, Google/Bing have it answering user questions directly, it's commonly pitched as a "tutor", or an "assistant", the OpenAI API is being shoved everywhere under the sun for anything you can imagine for all kinds of tasks, and nobody is attempting to clarify it's weaknesses in their marketing.

As it becomes more and more common, more and more users who don't understand it's fundamentally incapable of reliably doing these things will crop up.

30% / 0% / 70%

Dang, this full out fooled me. Concerning, I guess we're here now. Lots to catch once you're aware of it, but totally passed by me while scrolling, even as someone who's well aware of AI Image Generation, even in an image that's intentionally ridiculous.

Honestly, makes sense, the active voice version is just... more efficient and easier to parse quickly.

Yeah, this is the problem with frankensteining two systems together. Giving an LLM a prompt, and giving it a module that can interpret images for it, leads to this.

The image parser goes "a crossword, with the following hints", when what the AI needs to do the job is an actual understanding of the grid. If one singular system understood both images and text, it could hypothetically understand the task well enough to fetch the information it needed from the image. But LLMs aren't really an approach to any true "intelligence", so they'll forever be unable to do that as one piece.